An Enhanced Medical Home for High-Risk Chronically Ill Children

INTRODUCTION

Although the patient-centered medical home is widely recommended to promote cost-effective healthcare1-6, it has not been shown in systematic reviews to improve clinical outcomes or reduce medical costs.3,4 Medical homes have greatest potential value for high-risk patients, including chronically ill children whose care is often fragmented, costly, and ineffective. However, the justifiably stringent requirements for practice guidelines and the necessity to control health care costs require compelling evidence of cost-effectiveness before undertaking all that will be required to disseminate the patient-centered medical home for high-risk chronically ill children throughout the U.S. For this reason we propose a randomized trial to assess whether an enhanced medical home providing comprehensive care is cost-effective in preventing serious illness (death, pediatric intensive care admission, or hospital stay >7d) among high-risk chronically ill children in our center. Our study obtained expedited approval by our Institutional Review Board (IRB) as a quality improvement (QI) study.

METHODS

After verbal informed parental consent is obtained (as allowed by our institutional review board for a minimal-risk QI trial to increase access to care), children are randomized (using sealed, opaque, sequentially numbered envelopes) to usual care or comprehensive care after stratification by maternal education (high school graduate or not) and predicted risk of hospitalization (50-74%; >75%).

Comprehensive Care involves care for acute and chronic problems from an ethnically diverse team of pediatricians and pediatric nurse practitioners (PNPs) who are highly trained and experienced in treating these complex, fragile children and available at all hours by phone and 40 hours/week in a special high-risk children's clinic. This clinic serves as a novel medical home where both primary and specialty services are provided in the same clinic at the same visit. The clinic is also staffed by a nutritionist and social worker and attended monthly by a dedicated subspecialist in pediatric gastroenterology, in neurology, and in allergy/immunology. A pediatric infectious disease specialist helped develop measures to reduce, promptly diagnose, and effectively treat infections. These subspecialists are available by phone for consultation at all hours. Acute problems presenting before 5 pm are seen that day; those occurring on weekends or nights are seen the next weekday in the morning. At any hour an emergency department (ED) visit or hospitalization is needed, our staff discusses the child with the responsible MD and schedules prompt follow-up visits. Multiple measures, e.g., recording and staff review of phone calls; daily checks of ED and hospital logs; detailed review of all care before hospitalizations; parent surveys; and active input of our Parent Advisory Board, are used to promote highest quality of care.

Usual Care is provided in the offices of private pediatricians or our general pediatrics clinic staffed by faculty-supervised residents. Chronic problems are treated in our subspecialty clinics. After-hours calls from parents to our center are taken by faculty-supervised pediatric residents or faculty unlikely to know the child. Children referred to our emergency department had no automatic follow-up appointment.

Statistical analyses and stopping rules: Intent-to-treat analyses will be performed using Poisson regression models with robust standard error estimators (to account for within family correlation and estimate relative risk) fitted to the number of children with a serious illness, intensive care unit admission, hospitalization > 7 days, or death. Negative binomial regression models will investigate group differences in the total number of such outcomes. Models will be adjusted for baseline risk, maternal education, and length of follow-up. A p<0.05 is considered statistically significant. Bayesian analyses will be performed to estimate the probability of reduced serious illnesses and of reduced costs (assuming a neutral prior probability of relative risk = 1.0; 95% credible interval = 0.5-2.0 [encompassing the largest likely effect size for major outcomes observed in randomized trials]). We planned to enroll 400 patients to identify a one-third reduction in total patients who developed serious illness (alpha error = 0.05; power = 0.80; projected serious illness rate with usual care = 38%). Under predefined stopping rules, enrollment would cease whenever Bayesian analyses performed annually from the end of the second year identified a >95% probability that comprehensive care is cost-effective.

Economic evaluation: Hospital costs (including costs for observation stays) will be estimated from a health system perspective by multiplying hospital charges (obtained from hospital's claims data and Medicaid billing records) by department-specific cost-to-charge ratios specified in the hospital's annual Medicare Cost Report. Outpatient costs for usual care will be estimated using standard methods based on relative value units. Outpatient costs for comprehensive care will be estimated using total clinic expenditures to include costs for start-up, longer patient visits, extra (unbillable) services, and low patient to staff ratios not addressed by relative value units. Costs will be inflated to 2014 U.S. dollars based on the Consumer Price Index for medical services. Cost differences between treatment groups will be assessed using generalized linear models with log-link and gamma distribution, adjusting for maternal education, hospitalization risk, length of follow-up, and within-family correlation.

The investigators will consider the program to be cost-effective if it reduces the total children with a serious illness without increasing total clinic and hospital costs, reduced these costs without increasing the total children with a serious illness, or reduced both.